|
|
Research Reports |
| Title: | Analysis of the Relationship of Roadside Inspections on Large Truck Crashes |
| Authors: | Pan Lu, Brenda Lantz, Denver Tolliver, and Zijian Zheng |
| University: | North Dakota State University |
| Publication Date: | Apr 2019 |
| Report #: | MPC-19-381 |
| Project #: | MPC-475 |
| TRID #: | 01706819 |
| Keywords: | compliance, crash causes, crash risk forecasting, crash severity, data mining, fatalities, injury severity, inspection, motor carriers, truck crashes |
| Type: | Research Report – MPC Publications |
This research seeks to investigate crash severity predicting models and contributing factor explorations through the application of data mining models. There are 21 variables found to be associated with commercial truck injury severities. The importance analysis indicates the variable relative important levels for contribution. The top 11 variables account for more than 80% of injury forecasting. For property damage only, the most important variable is "Carrier State," which indicates that the variable of Carrier State makes the most contributions, as compared with the other variables in explaining property damage only crashes. Variables contribute differently when explaining different crash severities. A variable showing significant importance for a certain severity level may be less crucial for another. For instance, "Cargo Body Type" is the second most important factor for predicting fatality crashes, but is much less important for predicting property damage only crashes (severity=0). However, it is clear that Carrier State is the most influential factor for all severity levels. Marginal effects of important variables are conducted and summarized in the research.
Lu, Pan, Brenda Lantz, Denver Tolliver, and Zijian Zheng. Analysis of the Relationship of Roadside Inspections on Large Truck Crashes, MPC-19-381. North Dakota State University - Upper Great Plains Transportation Institute, Fargo: Mountain-Plains Consortium, 2019.
View Report